Automated Low-Cost In Situ IR and NMR Spectroscopy Characterization of Clinical-Scale ¹²⁹Xe Spin-Exchange Optical Pumping

Abstract

We present on the utility of in situ nuclear magnetic resonance (NMR) and near-infrared (NIR) spectroscopic techniques for automated advanced analysis of the 129Xe hyperpolarization process during spin-exchange optical pumping (SEOP). The developed software protocol, written in the MATLAB programming language, facilitates detailed characterization of hyperpolarized contrast agent production efficiency based on determination of key performance indicators, including the maximum achievable 129Xe polarization, steady-state Rb–129Xe spin-exchange and 129Xe polarization build-up rates, 129Xe spin-relaxation rates, and estimates of steady-state Rb electron polarization. Mapping the dynamics of 129Xe polarization and relaxation as a function of SEOP temperature enables systematic optimization of the batch-mode SEOP process. The automated analysis of a typical experimental data set, encompassing ∼300 raw NMR and NIR spectra combined across six different SEOP temperatures, can be performed in under 5 min on a laptop computer. The protocol is designed to be robust in operation on any batch-mode SEOP hyperpolarizer device. In particular, we demonstrate the implementation of a combination of low-cost NIR and low-frequency NMR spectrometers (∼$1,100 and ∼$300 respectively, ca. 2020) for use in the described protocols. The demonstrated methodology will aid in the characterization of NMR hyperpolarization hardware in the context of SEOP and other hyperpolarization techniques for more robust and less expensive clinical production of HP 129Xe and other contrast agents.

Pilot Quality-Assurance Study of a Third-Generation Batch-Mode Clinical-Scale Automated Xenon-129 Hyperpolarizer

“…At the rear of the SEOP cell, a mirror is used to retro-reflect the laser beam to provide a second pass through the SEOP cell. NMR and NIR data were processed automatically using a MAT-LAB software package described previously [67]. This automated open-source software package is freely available in the electronic supplemental information of Ref.…”

“…This automated open-source software package is freely available in the electronic supplemental information of Ref. # [67].…”

“…Before measuring the HP 129 Xe build-up rate, the SEOP cell was heated to the desired set point while the laser current was maintained below the nominal operating current of 37.5 A, corresponding to an off-resonance condition, i.e., virtually no photon absorption occurs by Rb. This practice ensures effectively negligible polarization at the start of each build-up experiment (performed at full laser power and on D 1 resonance corresponding to 37.5 A current), which simplifies the fitting mono-exponential build-up function to y = %P Xe *(1-exp(-x/T b ), where T b is the build-up time constant (mins), x is time (mins), and %P Xe is the maximum P Xe at t → ∞ (note the fitting function has two variables as y(0) = 0), described in detail in previous work [67].…”

“…Here, T aq is the pre-acquisition delay time to mitigate the ring-down effect of the RF coil (3 ms in our experiments), and T 2 * Xe and T 2 * H are the observed T 2 * spin relaxation time constants in ms for 129 Xe and 1 H, respectively. At the end of each build-up experiment, an exponential curve fit of 129 Xe polarization build-up over time was performed using Equation (3) via automated MATLAB processing software described previously [67]:…”

“…129 Xe T 1 under both configurations was comparable and appreciably long, i.e., in excess of two hours, Figure 7a,b, respectively. Xe T 1 as described previously [67] (see SI). Lower displays show comparisons of (c) 129 Xe T 1 values (note the y-axis starts at 40 min to better delineates the differences) and (d) 129 Xe polarization build-up curves recorded for the first four experimental points shown in display c-note the consistency in the polarization dynamics for a given cell configuration.…”